We Democrats message poorly on economics. This diary shows a way to help people understand a simple point on the rate of consumer price increases (inflation):
Inflation is already falling back toward pre-pandemic, pre-Russian invasion levels. Over the last three months, the consumer price index (CPI) hasn’t increased much at all.
This message is simple and accurate. See the grey line in the chart above (3 month annualized inflation) vs. the blue line (12 month inflation).
There are certainly caveats and complexities with any statistic.
It’s true that the “headline” inflation rate, which measures how much prices have increased over the last 12 months, is still 8 percent. Prices shot up in the first 5 months of 2022, mostly due to the Russian invasion of Ukraine. Thus, the headline 12-month inflation rate will remain high throughout 2022. But by mid-2023, after those war-related price increases have rolled off the 12 month growth rate, headline inflation will seem much more tame.
But I just prefer to use a shorter measurement period to analyze inflation. Annualizing short term data is very common in economic statistics. The math is simple: for example, if GDP goes up by a measured 1 percent over a quarter (three months), then BEA publishes a headline estimate that GDP grew by 4 percent that quarter at an annualized rate (1% times 4 quarters in a year). Likewise with prices, if the CPI actually goes up by 0.5 percent from one month to the next, that implies a 6 percent inflation rate on an annualized basis (0.5% times 12 months in a year).
There are certainly pros and cons. When prices are especially volatile, extrapolating one-month’s data to a full year could seem unrealistic or improbable. But to rely only on annual or 12 month backward looking growth rates tells us very little about what’s currently happening — it leaves us stuck mostly measuring the past.
Here’s an example: In both March and June of 2022, the monthly CPI price increase over the prior month was more than 1.3 percent, which would extrapolate to a 16 percent annual rate. That would seem an exaggeration to extrapolate from a volatile monthly bounce.
However, using 12 month backward rates instead leads to hyperventilating and misleading ledes like this particularly bad article from CNN, which made it seem like inflation was at an 8 percent annual rate in September. No, it wasn’t. The non-seasonally adjusted CPI actually increased by less than 3 percent (annualized) in September and the seasonally adjusted CPI actually increased by less than 5 percent (annualized).
Where that CNN story said
On an annual basis, prices rose by 8.2% in September, a slower increase than the 8.3% rise seen in August, according to the Consumer Price Index, which measures the changes in prices for a basket of consumer goods and services. [my emphasis]
It should have said
On an annual basis, [P]rices rose by 8.2% over the last 12 months (between October 2021 and September 2022), a slower increase than the 8.3% rise seen in as of August, according to the Consumer Price Index, which measures the changes in prices for a basket of consumer goods and services. [my rewrites in bold]
Here are some additional thoughts and context:
1. The 3 month annualized rates that I like to use for short-term inflation tracking don’t take into account seasonality. Many businesses raise prices in January, and if you look closely at the month-to-month price changes, you can see that the price spike I’ve “mostly” linked to the Russian invasion actually started in January, before worries about fuel and food prices started to get factored into prices. On balance, I think ignoring seasonality is fine for my purposes making the basic point about the difference between 12 month and 3 month annualized inflation rates in the current context.
2. CPI inflation is heavily influenced by food and fossil fuel prices, which are especially volatile. That is why many economists like to strip out food and carbon fuels, to look at “underlying” price increases. Absent food and fuel, the annualized 3-month inflation rate has dipped much less, down to about 5 percent in September (orange line in the chart adjacent). Of course, food and fuel prices are also still working their way through the costs and prices of other goods, and that process has a lag.
3. The Fed is very concerned with inflationary expectations — that is, even after the “shock” of higher food and fuel prices works through the economy, businesses and workers could preemptively demand higher prices and wages because they expect consumer prices to keep rising rapidly. In my opinion, inflationary expectations were a big deal in the 1970s, but aren’t so important now. Much smarter people than I are saying bond markets seem to expect inflation to be reduced back toward the Fed’s target of 2 percent quite quickly.
4. Price increases occur because demand is greater than supply. This year and last, I believe much of our price inflation has been due to inadequate supply rather than too much demand. A Fed researcher recently did a very cool study trying to sort out how much of our recent inflation (based mostly on pre-war data) was due to supply constraints vs. demand. (Hint, it’s mostly supply, but you should read the whole article — it’s very accessible.)
5. Lower inflation doesn’t necessarily mean lower prices. Inflation — the rate of growth of prices — has to go negative to get prices to actually fall. We sometimes call a reduction in the inflation rate “disinflation” when the rate comes down but doesn’t necessarily go negative.
6. I believe the lack of competition is also boosting prices. Some businesses seem to be able to raise prices even more than their costs have increased, presumably because there’s not much competition for their products or services right now. Hopefully, price competition will return as supply line constraints ease, but you can also make a case that many important businesses have essentially become monopolies, effectively blocking price competition. This is a subject for another diary.
Summary: No statistic is perfect, but we can truthfully and non-misleadingly tell voters that inflation has started heading down since early this summer. The context and complexities of measuring and predicting inflation are important, but using a shorter time frame seems much more helpful at assessing the inflation that’s happening now.
My math for this article (at least for main CPI):
CPI for All Urban Consumers (CPI-U) |
|
|
|
Not Seasonally Adjusted |
12 month growth |
Annualized |
Annualized |
Annualized |
|
Index |
rate |
6 month rate |
3 month rate |
1 month rate |
2020Jan |
257.971 |
2.5% |
1.1% |
1.0% |
4.7% |
Feb |
258.678 |
2.3% |
1.7% |
2.3% |
3.3% |
Mar |
258.115 |
1.5% |
1.1% |
1.8% |
-2.6% |
Apr |
256.389 |
0.3% |
-0.7% |
-2.5% |
-8.0% |
May |
256.394 |
0.1% |
-0.6% |
-3.5% |
0.0% |
Jun |
257.797 |
0.6% |
0.6% |
-0.5% |
6.6% |
Jul |
259.101 |
1.0% |
0.9% |
4.2% |
6.1% |
Aug |
259.918 |
1.3% |
1.0% |
5.5% |
3.8% |
Sep |
260.28 |
1.4% |
1.7% |
3.9% |
1.7% |
Oct |
260.388 |
1.2% |
3.1% |
2.0% |
0.5% |
Nov |
260.229 |
1.2% |
3.0% |
0.5% |
-0.7% |
Dec |
260.474 |
1.4% |
2.1% |
0.3% |
1.1% |
2021Jan |
261.582 |
1.4% |
1.9% |
1.8% |
5.1% |
Feb |
263.014 |
1.7% |
2.4% |
4.3% |
6.6% |
Mar |
264.877 |
2.6% |
3.5% |
6.8% |
8.5% |
Apr |
267.054 |
4.2% |
5.1% |
8.4% |
9.9% |
May |
269.195 |
5.0% |
6.9% |
9.4% |
9.6% |
Jun |
271.696 |
5.4% |
8.6% |
10.3% |
11.1% |
Jul |
273.003 |
5.4% |
8.7% |
8.9% |
5.8% |
Aug |
273.567 |
5.3% |
8.0% |
6.5% |
2.5% |
Sep |
274.31 |
5.4% |
7.1% |
3.8% |
3.3% |
Oct |
276.589 |
6.2% |
7.1% |
5.3% |
10.0% |
Nov |
277.948 |
6.8% |
6.5% |
6.4% |
5.9% |
Dec |
278.802 |
7.0% |
5.2% |
6.6% |
3.7% |
2022Jan |
281.148 |
7.5% |
6.0% |
6.6% |
10.1% |
Feb |
283.716 |
7.9% |
7.4% |
8.3% |
11.0% |
Mar |
287.504 |
8.5% |
9.6% |
12.5% |
16.0% |
Apr |
289.109 |
8.3% |
9.1% |
11.3% |
6.7% |
May |
292.296 |
8.6% |
10.3% |
12.1% |
13.2% |
Jun |
296.311 |
9.1% |
12.6% |
12.3% |
16.5% |
Jul |
296.276 |
8.5% |
10.8% |
9.9% |
-0.1% |
Aug |
296.171 |
8.3% |
8.8% |
5.3% |
-0.4% |
Sep |
296.808 |
8.2% |
6.5% |
0.7% |
2.6% |